Self-delusion is a winning survival strategy, study suggests

However, this bold approach also risks wreaking ever-greater havoc. The authors cite the 2008 financial crash and the 2003 Iraq war as just two examples of when extreme overconfidence backfired.

A team from the University of Edinburgh and the University of California, San Diego used a mathematical model to simulate the effects of overconfidence over generations. It pitted overconfident, accurate, and under-confident strategies against each other.

A paper published in Nature September 14 shows that overconfidence frequently brings rewards, as long as spoils of conflict are sufficiently large compared with the costs of competing for them. In contrast, people with unbiased, accurate perceptions usually fare worse.

The implications are that, over a long period of time the evolutionary principal of natural selection is likely to have favored a bias towards overconfidence. Therefore people with the mentality of someone like boxer Mohammad Ali would have left more descendents than those with the mindset of film maker Woody Allen.

The evolutionary model also showed that overconfidence becomes greatest in the face of high levels of uncertainty and risk. When we face unfamiliar enemies or new technologies, overconfidence becomes an even better strategy.

Dr Dominic Johnson, reader in Politics and International Relations at the University: "The model shows that overconfidence can plausibly evolve in wide range of environments, as well as the situations in which it will fail. The question now is how to channel human overconfidence so we can exploit its benefits while avoiding occasional disasters."